Why to use Load Balancer with Floating IP?
I can notice every tutorial for using Floating IP is explained through this schematic:
https://assets.digitalocean.com/articles/high_availability/ha-diagram-animated.gif
I'm trying to understand why we need to use Load Balancers in combination with Floating IP ? Why not to use only pool of App servers with Floating IP (without Load Banalcers) ? When some server is down, the floating IP will be assigned to other server from the pool and the application will be still available.
Thanks
I have this very same question as well. I don't understand the need for servers with HA Proxy installed. What if your app servers have a higher bw than the smaller low end servers with HA Proxy? The HA servers would reach their limit of bw faster.
Without Load Balancers, you cannot use multiple App servers.
Floating IP is not a dynamic switch mechanism (aka load balancer), instead it resolves biggest problem Digital Ocean had - lack of ability to have static yet dynamic IP's in DNS settings.
With it you don't need services like CloudFlare to manage DNS.
Unfortunately, there's a catch. One droplet can support just one Floating IP. It means you cannot replace broken server with other LB server assigned to other Floating IP. It means only way to have proper redundancy is to have idle spare Droplets . If you will use 3 Floating IP's you need to have 3 LB servers + at least one spare.